3 = 1 + 2: how the divide conquered de novo protein structure prediction and what is next?

作     者：Yaoqi Zhou Thomas Litfin Jian Zhan Yaoqi Zhou;Thomas Litfin;Jian Zhan

作者机构：Institute of Systems and Physical BiologyShenzhen Bay Laboratory Institute for Glycomics Griffith University 

出 版 物：《National Science Review》 (国家科学评论(英文版))

年 卷 期：2023年第10卷第12期

页      面：60-65页

核心收录：

学科分类：0710[理学-生物学] 071010[理学-生物化学与分子生物学] 081704[工学-应用化学] 07[理学] 08[工学] 0817[工学-化学工程与技术] 

基　　金：supported by the National Key Research and Development Program of China(2021YFF1200400) the Major Program of Shenzhen Bay Laboratory (S201101001) a Griffith University Postgraduate Fellowship 

摘      要：Protein structure prediction has been a challenging problem for60 years. Its progress has been considered incremental for the past 20 years until leaps were made by Alpha Fold that seemed to suddenly reach experimental accuracy. We argue that this unexpected success in 3D structure prediction was foreshadowed by prior advances in predicting 1D backbone structures and 2D side-chain distance maps in continuous angle and distance space as well as the embedding of these 1D and 2D features for end-to-end learning.